Streaming Audio Networks - What matters? And what doesn't? Noir & Blanc hifi show 2022
ฝัง
- เผยแพร่เมื่อ 6 พ.ย. 2022
- At the first edition of the Noir & Blanc Hifi Show in Brussels, we did a presentation about Streaming Audio Networks. We explained very shortly how networks work, the different media you can use to build a network and how to tweak it. And yes: we talked about the importance of decent equipment. Incluiding measurements to prove it :-).
- วิทยาศาสตร์และเทคโนโลยี
I've learned alot thanks to Alpha audio!!
My Asus router had to be restarted every 2 weeks or so, I changed the psu to an Sbooster and since never needed to restart my router. POWER MATTERS :-)
Well explained, good job! I'm a network carrier guy as well and never believed it...but, started in my HiFi chain with the Paul Pang Dual OCXO and played with Ferrum Hypsos on my streamer... big difference!! Changed PSU on my AVM fritzbox, another difference... thanks a lot
Great presentation Jaap and I have certainly seen much benefit to my digital front end by upgrading my switch to an English Electric one from Chord and using better network cables. Keep up the good work :-)
Thank you Jaap. Very educational and informative for me. I will certainly upgrade my modem/router and look at sourcing a decent network switch. Kind regards Rob Australia 😎👍
You are very welcome!
Interesting presentation & I look forward to more findings in this contentious area........ I don't mean to come across as suffering from "Obsessive Compulsive Disorder" & "nitpicking" but @ 26:08 you mention "fiberglass" in your overheads, the general convention is "optical fiber" or just "fibre" ( without mentioning, single mode or multimode or real glass versus plastic optical fiber ... ) the term "fiberglass" is reserved for applications such as a "fiberglass boat or surfboard" or fiberglass thermal or high temperature electrical insulation - I like everything you "guys" do & it's refreshing to hear an intelligent conversation on a subject that's dear to me ( sound quality ).....
You are absolutely correct... Fiberglass is something else... :-).
I found cascading 2 merakis with fiber and startech SFP's worked very well. I replaced the power supplies with LPS and a LC filter. BIG improvements
Cisco Meraki makes very nice switches. I can imagine the power supply upgrade did wonders.
Which Merakis switches are recommended? I’m seeing one model on Amazon for small offices at $209 and there are some Merakis that are thousands of dollars.
@@aa5az423 220
Great presentation! What is the best way to upgrade the powersupply of the meraki switch?
There are some upgrade kits available. Yeti acoustics has a mod.
Would love to see your review/testsof Audiophool's modified Cisco Meraki switch
That's so STUPID, another spin at companies and resellers like him trying to abuse consumers...
Thanks, great presentation. I do have a question though… should all cables leading to relevant audio equipment be connected to the same switch? For example, NAS, Roon Core, Roon endpoint. I have Roon Core and NAS in the switch but the Ethernet cable that leads to the Roon endpoint is plugged directly in the router (a rather cheap one). I do have a galvanic isolator at the end of that cable, in the connection to the endpoint. Thanks!!
No its not necessary to connect everything to the same switch.
Hi Jaap, thanks for this very informative video. Very nicely explained. Do you have any experience with the Innuos Phoenix NET switch? If so, what your take on that switch?
Best Regards:Kris
Thank you Jaap !
Are you saying that I can greatly improve performance by substituting the isp provided modem/router with a separate modem and separate router ?
Welcome... And yes... That is what i am saying.
Hi Jaap
Thanks again for jet another interesting video👍
I have designed and build networks professionally for over 16years.
I have built and used media servers in my HiFi setup for about 10years, only the last 2½years have I optimized my private network at home.
Here I also learned that network for HiFi does make a huge difference 👍😊
In your video did you state that Network speed does not matter, because bandwidth requirement for Audio is so low.
Yes, bandwidth requirement for Audio is low and like you said maybe around 10Mb/s when it is high.
Most switched and NIC's is about 100Mb/s or 1Gb/s, which is 10 or 100 times higher than the audio streams needs.
Still speed of the NIC's does matter, if you take a 1Gb/s NIC and lock it till 100Mb/s your sound will change.
If you use a managed switch, be sure to set speed and duplex to the same in both ends.
I will not say which one is best, because some systems will benefit for using full speed other systems will benefit if using 100Mb/s on a 1Gb/s NIC.
If you have tested this too or if you try it, please let me know what you think :)
Thanks
Br
Jacob Land Jensen
Denmark
Thanks for the insights. I will try it and write some stuff about it!
Thank you for this interesting Video. You talk about a DAC in your Video. Which one you talk about? THX
A good DAC doesn't cost much money, an SMSL SU-1 is already better than your ears and any analog gear!
The double shielding which you suggest improoves sq only on digital connections?
Thank you
John
DON'T USE STP!! (Shielded Twisted Pairs) It just brings more problems than what it is trying to solve! If you have enormous source of interference (like passing aside of a big 600V power tranformer), then use optical fiber, it's cheap !
Hello Jaap, what do you mean with shielding on both sides? What kind of shielding you used?
The cable shielding should be connected to both connectors.
Hi Jaap, going trough your presentation Have you ever tried to measure the voltage on the ethernetwire of a mediaconverter to find out its default output noise level ? I would expect that for -110 dB you need to be in the range of 10microvolt ?. Would it be feasible you think ?
We can definately do that. We will receive our CDN next week. I will finish the measurements after installation. I can check a converter than as well.
Complete waste of time, we're talking about a well defined and robust specification and the usual digital logical signal with detection levels... Microvolts don't have any significance in that... Think of a multi-gigahertz CPU with billions of gates arranged in multiple layers: Do you think you can have worst case than that? Still, CPUs don't make any errors!
If I put a good power supply to my wifi router, should I get same improvements for both cable and wifi connection to my streamer?
You won't get anything out of that, it's just plain waste of money... Telecom equipment are design for error-free transmission of data, even the cheapest ones...
That was great, you broke down all of the relevant issues in a very structured way that was easy to follow. I have an engineering background but all of this is foriegn territory to me. Thanks very much.
Welcome!
@@TheAlphaAudio Also Ecdesigns don’t need fancy streamers as long as it’s bit perfect .. even a laptop would do .. not easy to believe but give a listen you will be fr shock
The way this information was delivered was top-notch.
I salute you 🫡
It's just that it's plain false.
Opinions differ about whether or not to connect shielding at both ends of the utp cable and some services swear by one or the other, occasionally I am specifically asked to mount shielding at both ends but usually the other way around
Opinions can differ. But it is technically completely incorrect to connect the shield to one plug. You actually introduce more noise ánd the shield doesn't work. And IEEE specifically states that shielding should be connected to both connectors, if the cable is shielded.
@@TheAlphaAudio I trust what I hear, in an industrial environment I would not dare to connect on one side as the IEEE standards certainly apply. But as I wrote ... I apply what the customer wishes :)
@@fernandlambert4342 OK
It's not a matter of opinion, just DON'T USE STP !!! Use fiber if you even encounter an extreme interference situation such as passing ethernet cabling aside a big 600V power tranformer!
Hi again ..time has passed and we’re now on a very fast and healthy network for our wifi 😎
Now is there a quality streaming unit that uses wifi ..my other half won’t let me pass cable throughout the house ..and that’s probably a good thing 😅
Nothing special, just enough buffer to account for variable speed data stream. If you abuse your WiFi (which is CSMA/CD), then music will stop and restart when the input buffer will fill enough.
GM ☕️..yea, it’s me …again.
My biggest issue with my streaming of music ( Amazon Utra HD) is my incredible slow internet speed..in a semi rural area….so once we upgrade from 6MPS to something around 50MPS ..now we can do more around here.
6 Mbps is more than enough... It depends if you abuse it, like downloading large files at the same time. And you'd be much better to use regular CD quality streaming, there's no such thing as "Ultra HD Audio", it's just lower noise floor and CD quality is already much better than your ears... If the HD version of the same album is better, it just means that the source material is not the same (different mix, remaster, etc).
Very interesting video, thank you! I have a Bluesound Node 130 upgraded with PD Creative PSU and fed with ifi Power X Powersupply, a Vodafone Station Router with standard Powersupply. All cables are supra cat 8 cables. Would you recommend a network switch or is it better to not have a switch? If you recommend a switch, which would you buy max 100 Euro? Would you upgrade the Powersupply for the Router or isn't that necessary? Thank you!!
A good switch with nice power supply will definitely help.
@@TheAlphaAudio which would you recommend? When I wont spend more than 200 euros for the switch and the power supply? Thank you very much:)
@@paulmastermann2200 Netgear GS108E with a nice IFI power supply.
@@TheAlphaAudio thanks a lot! The Netgear is now in my shopping cart. Would the ifi ipower with 12 V and 1,8 Ampere (50 Euros) be a good match with the Netgear or would you recommend the if power X with 12 V and 2 Ampere (100 Euros)? Thanks :)
@@paulmastermann2200 you dont need a lot of current... Even 1A is enough.
@8:30 UTP cabling is CMNR (common mode noise rejection). You don't need shielded in home applications. T.I. has a paper on EMI effects on 10/100 Ethernet PHY's and the conclusion is noise immunity below 30MHz. That's the first thing to understand. On STP cable you need to look at the Siemens paper called the 'Antenna Myth' as you are now, unfortunately, perpetuating those same myths.
are you still using the Meraki ?
Yes
How i shield in two side in a cat 6a cable where already have a ground drain cable??? May the drain cable and shield foil connect together??
Yes. You can also check with a cheap cable tester.
@@TheAlphaAudio and the drain wire and foil crimped on the plug ground body right??
@@sndpds50 exactly
@@TheAlphaAudio wow. Then I’m the right path. Tnx for ur informative videos.
Will try for a network switch with my rpi4 definitely within few weeks.
@@sndpds50 thanks. We will be doing a mass test soon.
Audioquest Pearl, and other Audioquest ethernet cables, have the shield tied to ground at one end only. You can confirm this by connecting a continuity meter between the connectors.
I dont think the Pearl is shielded at all to be honest
@@TheAlphaAudio ??? A Pearl cable is CAT7, which requires a shield and one was taken apart in the video to show its shielding.
@@audiobomber true... Sorry. Was confused.
Hello Jaap. Thank you for the excellent presentation. Really like the fact that you stress - from the beggining - what the problem is in networks...Not 0/1 but noise....
Is there any way to link the slides? Would be wonderful to see the measurements diagrams in detail. I think you also publish this on your web. Would details be there?
Also I am not exactly clear what power supply you propose as best to use with switch. A good linear or a good switching power supply?
Thanks!
Hi Magmamin,
Thanks! Glad you like it!.
These measurements are already online. You can find them in this article: alpha-audio.net/background/why-a-good-switch-does-matter-for-streaming-audio/?highlight=%22why+switch%22. The new measurements and findings are still being processed.
@@TheAlphaAudio Perfect....Thank you
Much enjoyed. Wish I could speak Dutch or any other language for that matter even a tenth as well as you can speak English!
Thank you! Much appreciated.
Mijn naam is Jaap en dit is mijn spreekbeurt.
😃
Zoiets ja ;-)
The name is.. Jitter where you try is there streaming songs with no jitter in...
the sea of jitter
No need from hi end wires, big money for switches the data coming from the source and on your streamer converting and..then... . Jitter so get your self a streamer, good network and go for fishing songs there are songs with no jitter.
Jitter has nothing to do with audio streaming... Another of those audiofool lies... Worst you can get is audio stopping if incoming data stream slow and buffer empty... The DAC clock is use to reconstruct the analog and random jitter is just noise... well below human ear capability... th-cam.com/video/TT9JL2yaIOA/w-d-xo.html
11:42 it's inherent in the the whole ethernet protocol prevents data errors
Wrong - the ethernet protocol allows to detect data errors but it does not take any actions to correct that errors. Thats why TCP/IP is used which has additional procedures for requesting packets with data errors again before the application gets them
Maybe you should watch to the end... the problem is not TCP/IP or data errors.
@@TheAlphaAudio You should read my post again - you said that the whole ethernet protocol prevents data errors - and this is not true. Ethernet does not ensure that the transmitted data is correct.
The problem with your video that you assume that noise in a switch (or on a Ethernet cable) does influence the output of the DAC. To proof this you need to measure the output of the DAC when used with different switches.
Noise in the analog world is different from noise in digital signal processing.
In the analog world once noise is added to the signal it could not be removed anymore.
And as each analog component adds some additional noise you have to reduce noise in each component as much as possible
In the digital world noise upto some degree does not matter
If you define that a logical "1" is represented by 5 volt and a logical "0" by 0 volt than adding noise of +/- 1 volt does not matter if the receiver treats everything below 2.5 volt as "0" and above 2.5 volt as "1".
And each digital component (down to a logical gate) "refreshes" the signal.
If an inverter receives a voltage of 2 volt it will treat it as a "0" - invert it to a "1" and sends that out as 5 volt (+/- its own little noise)
So instead as in the analog world where noise is getting greater with every processing step in the digital world the signal is cleaned at every step.
When streaming music the data (from Internet) which is send to the streamer goes through thousands of digital processing steps (first it is processed by the TCP/IP stack, when it is normally decrypted and decompressed and after that it goes to a digital buffer - so a lot of "cleaning" takes place with the result that a little noise from the input of the streamer will not flow to the DAC.
This cleaning in each component is the reason why you can send gigabytes of data arround the world with 99,9999% chance of no errors. And in the very rare case that an error occur TCP/IP will handle it.
Digital to Analog converters are not only used for HiFi - they have a lot of industrial and scientific applications but none of that users has any of the "problems" so called "audiophiles" believe to have.
And DACs in HiFi only have to handle very low data rates and output frequencies compared to many other applications.
@ 13:32 "Are there any Electrical Engineers in the ... [ audience ] ... no? Ok, I'm not one as well, but... "
... and it shows. If there were any EE's in the audience, they would have contested these claims easily. In the digital domain, the signal on the cable is being interpreted as a 0 / "zero" if below threshold, and a 1 / "one", if above threshold. TTL logic thresholds are set such that there is a small gap in voltage values between the two states. For example, from 0 volts to 0.8 volts is considered a “low” logic state (zero / 0), and 2 volts to 5 volts is considered a “high” logic state (one / 1). Anything in between 0.8 to 2 volts is not considered a "legal" TTL logic value.
What this means is that although network and other data cables (e.g. USB, etc...) can and will be subjected to noise from external sources, generally in a very high percentage of cases this noise falls well below the threshold required to flip a bit on the wire. If that were to happen, you would experience data or packet loss (would be measurable and provable), and the end result (with streaming audio gear) might be some sort of audio drop-out, buffer underrun, or in worst cases playing a corrupted PCM data stream. However, in practice this is _very_ rare to see with well-designed gear (e.g. streaming devices, DACs, audio interfaces, etc...)
When packets arrive at the destination, they are processed by the receiving device and eventually translated into analog audio, usually by a DAC. If you were to test this DAC output for noise and compare between pieces of network gear in between, you'd see that most all of the noise floor is the same across tests and the entire frequency range. The only exception is if you hooked up the final DAC, amp and/or speakers in such a configuration as to create a ground loop. That would be considered a wiring error, and would immediately discredit any results of testing the two pieces of networking gear, unless one of the two pieces of gear indeed is somehow contributing to the ground loop in the system (in which case, that piece of gear would be considered faulty & bad design).
When you're comparing the two network switches and supposedly "hearing a difference", that is easily the placebo effect. Unless you're doing double-blind statistically interpreted experiments on this, you can easily fool yourself into thinking that you're hearing a difference. In reality, we often can "hear a difference" between two sets of gear or types of interconnections simply because we're focusing on the minutia of hearing something, and our brain and cognitive biases trick us into thinking we hear something as better or worse than the other identical audio when in reality it's the same.
Yeah, very basic principles, undergrad level in Electrical Engineering... Streaming audio must use buffering. TCP/IP is very robust and transparent even with errors (TCP checksumming and retransmission). You won't get corrupted PCM unless you have a broken audio device / streamer. And you're 100% right about placebo th-cam.com/video/BYTlN6wjcvQ/w-d-xo.html
network streamer is nothing but an ordinary PC with a better sound card packed in a branded box and sold at an absurdly high price
Uhhhh... No
@@TheAlphaAudio Why your "no", that's exactly that!
So this is why CD's still sound better most of the times?
Depends...
@@TheAlphaAudioIn HiFi …the answer is always “it depends”
If you stream the same CD quality data, there's no difference.
Play a wave file extracted from a cd sounds better then any streamer .
Why buy one its all polluted with wifi signals .
There's no such thing as "polluted by WiFi signals"... The audio quality of a streamer depends on the quality of the built-in DAC. It can be better or worst depending on which standalone DAC you compare it to.
Guys full of it, he's a salesman. Let's see him do a controlled double blind test, he claims he can hear differences in cables 😂 Why is he adding a switch into his setup anyway? Keep the chain simple ffs.
You won't see him, he'll never put himself in this situation and make a fool of himself... That's just regular peddler stuff here.